1 Association Rules - University of Melbourne（1关联规则-墨尔本大学）.pdf

Association Rules (Market Basket Analysis) Association rules are of the form if X then Y. For example: 60% of those who buy comprehensive motor insurance also buy health insurance; 80% of those who buy books on-line also buy music on-line; 50% of those who have high blood pressure and are overweight have high cholesterol. These rules are actionable in that they can be used to target customers for marketing, or for product placing, or more generally to inform decision making. Examples of areas in which association rules have been used include • Credit card transactions: items purchased by credit card give insight into other products the customer is likely to purchase. • Supermarket purchases: common combinations of products can be used to inform product placement on supermarket shelves. • Telecommunication product purchases: commonly associated options (call waiting, caller display, etc) help determine how to structure product bundles which maximise revenue • Banking services: the patterns of services used by retail customers are used to identify other services they may wish to purchase. • Insurance claims: unusual combinations of insurance claims can be a sign of fraud. • Medical patient histories: certain combinations of conditions can indicate increased risk of various complications. We consider how to derive association rules directly from historical data, as opposed to via customer surveys or other means. Such data is characterised by being readily available in large quantities (and thus cheap), though it is often of poor quality or incomplete. Theoretical Framework We will discuss the problem in the context of supermarket purchases, which is where the terminology “market basket analysis” comes from. The data consists of a number of transaction records, each containing a set of items purchased by that customer. For example Customer